Roman's 4GR V6 Toyota Carina

[Roman]

Since I've been nerding of the port shape stuff a bit, and have had a variety of interesting feedback on it. 
Including a few people saying that the current cross sectional area is plenty. 
I decided to see what sort of papers Toyota had about the development of the engine. 

On the JSAE website:
https://tech.jsae.or.jp/paperinfo/en/

You can do a search and then it was only 11 bucks or so, to download some articles from Toyota.
Pretty cool, although, needed translation. 
Most of the 4GR/3GR stuff was moderately interesting but not overly applicable. 

I always assumed that DI engines fairly heavily leaned on a stratified charge mode to get good economy. 
One of the articles said that these motors are designed for homogenous charge nearly 100% of the time.  
That's the only time it ever uses stratified charge is on cold starts to warm up the catalytic converters extra quickly.

But there were a few cool diagrams, this one is about what sort of compression ratio an engine can run, based on bore size and type of gas (from an OEM perspective)
Interesting to see that bore size is such a factor, I would not have guessed.
But also goes to show that direct injection is a huge advantage for knock resistance.

 

They had some cross sections from their CFD testing when developing the engine. 
Looks as though the purpose of the naca duct geometry at the port is to encourage more air to flow fully around the back side of the valve as well.



So probably a good idea to just keep it there.
Then this was interesting comparing the amount of friction gained/lost from the predecessor engine (2JZFSE) 
Interesting that it gains efficiencies in a lot of places but then loses a bunch again, to "other" which is probably stuff relating to having 2x banks instead of 1.




Anyway, so keeping all of this in mind / disregarding it all.



I've ordered a hand reamer to try and enlarge the ports anyway on a spare set of heads haha. 
There's a long enough straight section, that is near enough to a circle already that I think it'll be interesting to try.
Because, why not? The whole point of using these cheap, garbage can engines is to not feel bad ruining testing them 


I've got my motor back in the hole, but not fully back together yet. 
I think I will cry tears of joy if my gearbox shifts gear properly now.
Being unable to change gears properly has been my biggest remaining issue to figure out.

[Roman]

A few things happening. 

Firstly

HOLY SHIT THE GEARBOX WORKS NOW 
Absolutely stoked that it was just issues with the clutch line!

I mainly have the goal of getting to the OS drags with a good setup for once. 
Dyno on the back burner for now but hopefully reroll before drags.

Basic plan is: 
-Hoosiers
-Traction brackets
-Better rear shocks
-hopefully launch control / flat shifting (need speedo sensor working well, and position sensor on clutch)

I pulled the rear shocks out and measured them, and found that the AE86 Cusco 14 way adjustables were a good length, and will fit. 
I bought a set, and they fit up great! 



At the same-ish time I also got some nangkang CRS fitted, so my 15+ year old R888 are finally in the bin. haha
The car understeers quite a bit now, and its very wallowy when I have rear shocks on the softest settings. 
But drives a lot nicer over bumps. So I'll start creeping up the rear shock stiffness until it seems good.

In some mind blowing sort of news. I've been doing a bit more work and testing on my program that puts GPS log data directly into my ECU logs. 
This showed something interesting. My engine rpm wasnt matching the GPS speed that it should, with the 4.3 final drive that I've had fitted for the last forever.
So I check on the blocklayer website, and test a few sample points of GPS speed vs rpm in each gear to see what it should be.
Far out! The numbers all match up only if I put in that the car has a 3.9, not 4.3.
So that must be what it actually has. 
3.9!
Bloody disgusting. haha.
This is going to be wildly wildly wildly more fun with a 4.7 or 5.3 final drive. good lord.

Anyway, one reason this is particularly interesting, is that the virtual dyno program which I'm quite fond of.
It works out the power figure based on only rpm from the logs. Then uses gear ratio, tyre size, final drive, weight etc to work out the power figure. 
So this has been reading incorrectly the whole time, because I've had the wrong final drive entered.
With the correct final drive in, its more like this:





If I look at my VE values from my old intake, and how much power it made. 
Then add a directly proportional torque gain for extra airflow, (probably a little bit optimistic)
Then I get a very similar result which is promising. 



If these numbers seem like an impossible jump, keep in mind just how much extra fuel its needed now. the difference is crazy 



I'm now super fizzing to get back to the dyno to find out for sure.
Anything over 250whp and I'd be happy. 
Hopefully I can get back up there soonish. Just keep having stupid things pop up.

[Roman]

I added an extra page to my link ECU datalog viewer program so that it can align a video to a log file, and then write some overlays on it. 
Or you can just view it in real time, its cool for something like seeing where an oil pressure dip happens exactly. 
Or check if your driving is derpy. I didnt actually go to the track though, because that would be dumb.




I've been trundling the car up the road and back a bit, trying to get its road manners a bit nicer. Has a bit of a stumble and hesitation on first throttle so I'm looking at ways to try improve the situation.
One thing that I've found is that sometimes when throttle is closed its trying to apply a negative duty to hold it shut. 
So then when you first come on the gas, it has to go through a fair bit of PID movement to get to a positive duty to open the throttle. 
So I've increased the minimum close amount by a little bit, and it's a lot better on first touch of the gas. 

I think I still have some degree of problem with the throttles opening slightly unevenly relative to each other, and I dont think I'll ever fully get on top of that. 
So I am going to try use some cam timing tricks to dethrottle the engine a bit so it likes to stay at higher throttle openings with less power. 
But hopefully some good economy at the same time. 

On a baseline drive up and down the road, fuel economy is quite horrible, sitting around 13something liters per 100km.

My canbus based dual wideband trimming works pretty good though, it doesnt oscillate/overshoot/etc like the factory implementation does. (cyl 1 trim trims bank 1, and same for 2&2)



One thing is I struggle to mentally envision what is happening with the cam duration and overlap and so on. Then how this changes as the car travels. 
So I figure it might be handy to add valve timing graph into my log viewer program so I can view some stuff about what the engine is actually doing and see it change in real time.
I've got the cam lift/duration mapped out and the cam centerlines positioned correctly to account for my non standard timing.

I can hit the play button and it will run through the log and show how the cams over around, ignition timing moves, injector timing and duty changes, etc. 

One thing I noticed straight away is that my injector timing was sitting smack in the middle of the overlap period when I had 360 deg end of ignition timing.
So if I move around the cams to worst case scenario for biggest overlap, and then use the slider to move the injector timing to where it never occurs during overlap. then about 400-425 degrees is always out of that zone.





At low injector pulsewidths and low rpm I might be able to fit the injection window on the other side of overlap so I'm not losing any fuel straight out the exhaust. 




One thing of note. Is that cam charts always show the horizontal scale as being degrees of crank rotation. 
But it looks really different if you change the scale to mm of piston travel. Because the piston dwells and TDC and BDC then accelerates through the middle. 
So the same graph as above but the scale changed, looks way different. It makes the overlap period and time spent past TDC/BDC look a lot less drastic.
As there isnt necessarily any airflow movement at those points, apart from effects of harmonics and inertia and so on.




I am expecting to see some trends where if you view the amount of overlap or atkinson in milliseconds rather than degrees. it might end up being a nearly static value. 
So then it could map out what cam timing you'd need (in degrees)  to reach that target.

Now all of this is interesting (grading on a Dave scale) but how does it help improve fuel economy? 
Dunno yet.
I think I still need to do some real life testing at low load and see what works best. 
But I do like the idea of moving the injection window to the other side of overlap at low load.

Will be interesting to see if it helps at all. 

[Roman]

I tried a few variations of cam timing to optimize economy, and interestingly it all seemed about same same really. 
So I set it to max intake advance, and max exhaust retard which gives mega overlap as this makes the car kinda brap on decel. haha.
Seems to have plateaued at around 12l/100km. Not super awful but not great either.

Anyway, 
in os drag day related news. 
I have been wanting to copy the clutch position sensor setup that I had on the echo, for this car. 
However my previous setup mounted on the clutch fork, and this has a concentric slave cyl. 

So I have added one of these:

F9-3 Pin Brake Pedal Position Sensor 15192340 13583976 For Opel Astra GMC Buick Verano Pontiac




It's nice that it's got the slot, as it means it doesnt need to pivot on the same point. So this is mounted to clutch pedal and slides on a cut down bolt.



Results...
Ever since swapping this motor in, I've been battling to get even just "okay" gear shifts.
Now though? 

HOLY SHIT THIS IS GOOD